Last 1/3 Semester Flashcards
how movement occurs on the basilar membrane
movement increases from base to apex up to the point of maximum displacement, then decreases rapidly as the wave travels up the cochlea a bit further
masking basic definition
the process where one signal is made less audible because of another signal that is either present at the same time or close in time
concepts of tone-on-tone masking
- when the probe tone and masker are similar in frequency, the probe tone is masked efficiently
- when the probe frequency is further away from the masker frequency, there is less masking and often no masking if the probe frequency is lower than the masker (it is difficult to impossible for the masker tone to mask probe frequencies that are lower than the masker frequency because of the physical characteristics of the basilar membrane and traveling wave)
stopped at 10/31
start on 11/5 (tone-on-tone)
LPC
level per cycle; the intensity present at each individual frequency
equation to find overall level of noise
OA=LPC + dB BW
the level per cycle plus the bandwidth expressed as dB
*find dB BW by taking 10log (the bandwidth)
critical bands for auditory filters
the range of frequencies most important in masking a pure-tone probe signal
*determine the width of the CB by finding the narrowest band that just masks the probe
*in general CB is about 1/4-1/3 octave
(wider in high frequencies and narrower in low)
is masking a probe with a narrow band of noise a linear function?
yes, if the probe is increased 10dB you can mask it by adding 10dB more to the masker
bandwidth of NBN used clinically
1/3-1/2 octave wide
number of critical band in the human ear
24-25
another name for critical band
bark
sizes of critical bands
vary in size with frequency; are roughly equal below 1000Hz, but then increase above 1000Hz
equivalent rectangular bandwidths (ERB)
frequencies in a NBN closest to the probe frequency are most effective for masking; not all frequencies are important for masking a tone; (frequencies within a CB are close enough in frequency to mask each other; masking a higher frequency sound takes a wider range of frequencies)
- rectangle superimposed on figure gives a convinient way to describe the limits of a filter
- the area under the rectangle is the same as the area under the curved auditory filter
- as intensity increases, the ERB widens (more frequencies important in masking)
relationship between critical bands and loudness
if tones are in different CBs and are presented together, more neurons will fire (because neurons from both CBs) and the sound will seem louder
**4 tones within one CB will seem softer than 3 tones in 1 CB and 1 in another
why are acoustic reflexes lower for BBN than pure tones?
BBN is perceived louder because it activated many CBs (used in SPAR testing to predict sensitivity)
remote masking
high frequency sounds masking low frequencies; not really allowed by basilar membrane, however there is a small amount of this kind of masking
possible explanations for remote masking
- acoustic reflex= high frequencies trigger the reflex which stiffens the chain inhibiting low frequency transmission
- efferent system= path of neurons inhibits function of cochlea and efferent system suppresses cochlear amplifier, helps w/ speech in presence of noise
- distortions= create even lower tone to mask the low tone (F2-F1)
- **masking is more than just result of waves on the basilar membrane, the CNS and OHCs are also involved
psychophysical tuning curves (PTCs)
a probe tone is presented
a narrow band masker is presented
as the narrow band masker gets further away it takes higher intensity to mask the tone
this correlates to the traveling wave theory
tuning curves for low vs high frequencies
curves for lower frequencies are wider than those for high frequencies because of the octave scale (non-linear) if they were on a linear scale it would make them all look similar (high frequencies are collapsed due to the scale)
Q10
a measure of sharpness (q= quality factor) gives you an idea of how good cochlea is doing with certain frequencies
- find the point of the tip of the tuning curve, and go up 10dB
- draw a horizontal line at this intensity, and find the frequencies that intersect the line
- —-take the point at 10 minus the other point at 10 and divide the frequency of the pure tone by this range
- the probe tone frequency divided by the range of frequencies that Q10 line intersected gives you the Q10 of the psychophysical tuning curve
- ***Q10 addresses the difference on the x-axis, Q10 gives ratio to compare
Q10 examples and values
ideally we want Q10 to be around 10 and not too wide (smaller than 3 and shows cochlear problem)
*normal range is 3-10
above 10 is narrow which just shows good discrimination
*higher frequencies are more narrow than low?
TEN test
threshold equalizing noise test (Moore)
(HL)= phones
(ER3)= inserts
broadband signal with a spectrum shaped so that it will elevate thresholds of normal hearers to the same level at all frequencies
TEN procedure
- hearing loss< or = 60dB HL, TEN is presented at 70dB HL
- loss > or = 65dB HL, TEN is presented at 10dB above threshold
- *if ten is uncomfortable loud or threshold is at 90dB HL (max for ten) then set it to the audiometric threshold
- if threshold is above 90 dB HL, can’t test
- present masker and find threshold using 2 dB steps
criteria for dead region with TEN test
1) a shift > or = to 10 dB from original threshold
AND
2) a shift > or = to 10dB above masking level
***indicates cochlear dead region
**if change is 5dB above noise, increase noise and repeat
**not used very much in US (helps to not amplify areas that can’t use it)
temporal integration
how the auditory system takes advantage of longer duration signals
*above 200-250msec is infinitely long
temporal processing and pitch perception
- a 3000Hz tone can be differentiated from a 3009Hz tone
- 1 microsecond change in period is what we can perceive
- tonality perception= perceive as pure tones if they are long enough, but as a click if they are short because limits of the transducer not allowing the intensity to have the gradual rise and fall it needs to get to the maximum without distortion, going to maximum without rise time will result in distortions creating a Broadband signal and doesnt have enough cycles to detect as a pure tone
gap types and definitions
- within channel= with pure tones; signal before and after pause is the same
- between channel= white noise; signal before pause is different than after
- ***these pauses can be really really short
gap detection and frequency
can detect smaller gaps for high vs low frequncy
- reason for this relates to auditory filters; the narrower the filter, the longer it vibrates
- *high frequencies has wider filters than low
gaps in white noise
behave like high frequency sounds
*2 msec gaps can be detected if the noise contains energy up to 12kHz
minimum starting time between 2 clicks we can detect
1-2msec
- yet we can’t tell which came first, need 20 msec to do that
- detection of gaps and which signal arrives first is important for speech processing and relates to distinguishing voiced from unvoiced consonants
- *abnormal timing detection can relate to speech perception problems
temporal modulation transfer functions
- the amplitudes of signals are modulated at different rates
- -if the rate is very slow, we will be able to detect small modulation changes
- -the faster the rate of modulation, needs larger modulation to detect
- –in case of modulations above 1000x/sec (1/milisecond), the amplitude fluctuation is not noted at all
name the two types of temporal masking
forward masking
backward masking
forward masking
present noise, turn it off, then present tone
- using a click as stimulus and white noise as a masker at a constant level
- -a delay of 1msec can have up to 40dB of threshold elevation
- -with 50 msec of separation, the masking effect is down to 10dB
- -by 200-250msec, there is no effect